Technical Intern (AI)
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
TL;DR
Technical Intern (AI): Contributing to the development and optimization of foundational technologies for an RLOps platform with an accent on LLM stack internals and applied machine learning. Focus on building distributed training components, executing empirical research experiments, and improving production-grade AI system reliability.
Location: Must be based in or able to work from the Paris or New York office (In-person)
Company
is a frontier AI startup building an RLOps platform to specialize and deploy LLMs into production for enterprise clients.
What you will do
- Develop and maintain internal LLM stack components using Python and Rust.
- Assist in profiling and optimizing GPU inference performance and latency.
- Execute systematic experiments to analyze model behavior such as DPO or PPO tuning.
- Build and manage data pipelines for reinforcement learning processes.
- Debug issues within distributed systems and ML-heavy infrastructures.
- Collaborate with engineers and researchers to document findings and iterate on AI models.
Requirements
- Must be in final year of or recently completed a Master’s degree in CS, engineering, or related field.
- Comfortable programming in Python.
- Strong interest in machine learning, AI systems, or large language models.
- Demonstrated curiosity and proactivity in fast-paced environments.
- Must be available for a 6-month in-person internship.
Nice to have
- Coursework or projects in distributed systems or systems programming.
- Experience with PyTorch, JAX, or similar frameworks.
- Background in research projects or open-source contributions.
Culture & Benefits
- Paid internship program.
- Direct mentorship from senior engineers and researchers.
- Exposure to real-world, production-ready AI systems at scale.
- Collaborative, research-driven team environment.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →